Home/Virtual Labs

Desktop PC
  • CPU: i5-2500K  (@3,30 GHz)
  • RAM: 8 GB
  • SSD: 128 GB
  • HDD: 7 TB
  • OS: Windows 7 & Linux Mint (all 64-bit)
  • Monitor: Dell U2412M

As you can see, it's a nice upgrade from my old desktop. I was able to run the following GNS3 topology without any issue, besides the 16 putty windows (i prefer to have many windows concurrently on screen instead of tabbed ones, because it helps viewing simultaneously what is happening on the network and copy-paste is faster).



Dynamips/GNS3
  • GNS3 v0.8.6
  • INE's 10 routers + 2 switches topology
  • 4 extra routers as CEs

I tweaked a little bit the INE's GNS3 topology in order to add more routers and create more complex scenarios. Also i had to make an xls with device/port mappings between Rack Rental and GNS3 topologies, because all full scale labs in INE's workbook do not include any GNS3 topology or initial configs (probably due to pushing people to rack rentals). At the same time, all IOS-XR routers in INE's GNS3 initial-configs have the original IOS-XR config (which is not applicable to GNS3), while all switches have a different config (probably from another model). All of these had to be corrected (which proved to be quite annoying) in order to be applicable to GNS3 topologies.



Routers/Switches

  • ME-3400E
  • 7200/NPE-G2
  • 7600/RSP720
  • ASR9k/RSP440
  • CRS/PRP

The above devices were available at my work (to resemble https://learningnetwork.cisco.com/docs/DOC-10121 as much as possible), but i didn't use them extensively. Just to verify some IOS-XR and ME-3400 commands which were not available on GNS3.



This is the page where i also kept track of all my lab ratings.

Please keep in mind that these are my own ratings (after comparing my answers vs the actual/correct ones), since i couldn't find any online SP labs that offer a rating (like CCIE R&S Mock Labs). Initially i was planing to do more of them, but judging from my weekly progress i decided to skip some of them.



DateLabEnvironmentDurationRatingComments
14/Dec/2013INE SP Full Scale Lab #1GNS311:3069/100Besides some silly mistakes, Multicast BGP was the only one i couldn't actually think of. More than 2 hours were spent on fixing INE's GNS3 topology.
21/Dec/2013INE SP Full Scale Lab #2GNS310:3072/100Some tasks were described in a strange way, something that in the real lab would require the proctor's help.
A VPNv4 task (that i tried to do using various hacks) had a solution using static routes, although it was not allowed (-7). For a 6VPE task i decided to use IPv6 over IPv4 BGP which led to an interesting bug about IPv6 next-hop connectivity (-7). In a MPLS-TE task i missed a simple command, so MPLS TE was misbehaving (-7). In a MVPN task i missed the RP setup on the PEs (-7).
26/Dec/2013INE SP Full Scale Lab #3GNS39:3082/100Some tasks used static mroutes, which was referenced in the initial instructions but i hadn't seen it; nevertheless i used them, but in the real lab that would require more careful reading of the instructions.
In an Inter-AS task i changed the IGP config instead of BGP (something that was not forbidden), but the result was the same (-6). In an Inter-AS/CsC task it took me 3 hours to find the non-working reason, the "next-hop-self" which wasn't set (-6). In a MVPN task i didn't add static mroutes for RPF fix, although multicast was working fine (-6).
3/Jan/2013INE SP Full Scale Lab #4GNS313:3082/100That was one hell of a lab! 4 tasks had an error in their instructions, but after a little bit of processing it became obvious; nevertheless i spent extra time on finding and fixing those.
In an OSPF task i missed the LSA optimization hint and went for LSA flood-reduction only (-3). In a Inter-AS task i couldn't find a way to send labels (-4). In a simple PE-CE BGP task i missed to advertise the PE-CE links into BGP (-3). In a PE-CE OSPF ask i couldn't find a way to extend area 0 (-4). In a MVPN task i didn't add a static mroute, because everything was working fine (-4).

12/Jan/2014INE SP Full Scale Lab #1Rack Rentals06:1587/100Although every task was working according to the instructions, i want to be as strict as possible with my personal rating, so i missed a task where i didn't enable frame-relay switching and clock-rate (-6) and on another task having to do with mVPN i had to look at the online documentation because i couldn't remember an IOS-XR command (-7).
It took me around 5 hours to finish the lab and another 1 hour to double verify everything.
19/Jan/2014INE SP Full Scale Lab #2Rack Rentals06:4583/100Although everything was working fine, i missed a L2 task where i allowed extra vlans in a trunk (-5), another L2 task where i missed to configure portfast on some ports (-5) and a mVPN task where i didn't fix an RPF issue although the instructions didn't include such testing (-7). Some missed tasks were the same like on my first try, which is annoying.
I took me around 6 hours to finish the lab and another 45 minutes to verify everything.
25/Jan/2014INE SP Full Scale Lab #3Rack Rentals07:1594/100Besides a silly mistake where i didn't enable multicast on the switches (-6), everything else went fine. Time management needs to be improved further.
26/Jan/2014INE SP Full Scale Lab #4Rack Rentals09:3088/100As it proved, running the lab on real routers (rather than on GNS3) makes things a lot more interesting, especially on IOS-XR devices.
So, in this lab, i missed a BGP task where i had to replicate the ipv4 session settings to the new  ipv4-labeled session (-4) and that broke a previous TE task (-4), and another VPNv6 task where i didn't configure all the peerings (-4).
If i had 8 hours to finish the lab (verification included), then i would have scored 73/100.

The labs that i was running on GNS3 sometimes had a different solution from the one that was given in INE's workbook, because GNS3 cannot emulate IOS-XR routers. This means that an IOS-XR quirk might not be applicable in GNS3 and vice versa. Most of the times, when my solution didn't account for IOS-XR specifics (either by actual config or by config notes), i considered the task points as lost, because i'was actually trying to think in actual lab terms rather than in GNS3 terms.

As already noted in a previous post of mine, my primary focus on the beginning was on giving correct and justified answers (even if that was taking many hours) rather than on doing it as fast as possible. I managed to improve my time as more labs were done.

3 comments:

  1. Any reason you chose not to use IOU to study?

    ReplyDelete
  2. Hi Barry,

    I had explained my reasoning in http://ccie-in-2-months.blogspot.gr/2013/11/my-schedule.html.
    I'm quite familiar with dynamips/gns3 and since time was an important factor on my schedule, i decided to stick with the well-known environment and spend this limited time on the actual technologies.

    ReplyDelete
  3. Hi,

    I am having issue when i am running MPLS TE and then running MVPN, as TE tunnel points to same router where RP is hosted.

    ReplyDelete